I am having the same issue and its only on a MAC (FF, Safari). If I run the flex swf in parallels I have no memory issues.
I submitted a bug # SDK-20625 --- In flexcoders@yahoogroups.com, Alex Harui <aha...@...> wrote: > > One workaround would be to implement the panel in actionscript using custom > layout and positioning logic, thereby avoiding constraints. That would prove > that those aspects of the popup are causing the problem. > > Alex Harui > Flex SDK Developer > Adobe Systems Inc.<http://www.adobe.com/> > Blog: http://blogs.adobe.com/aharui > > From: flexcoders@yahoogroups.com [mailto:flexcod...@yahoogroups.com] On > Behalf Of Dmitri Girski > Sent: Monday, February 09, 2009 11:55 PM > To: flexcoders@yahoogroups.com > Subject: [flexcoders] Re: Memory consumption by the browser > > > Ryan, > > Well, it is a big popup, based on a customised Panel component, with 3 > states, declared in MXML. All elements are bound to the single data > structure which is passed as the parameter at the init stage. > > I am using my own caching PopUpManager, which tries to find the object > in the dictionary before using createPopUp() method. > > I know about the weak reference parameter in the event subscription, > but this is not a case - this panel uses mxml bindings, i.e. > click="fClickedThat()". > > As I said, I ran it under profiler and every time > PopUpManager.addPopup() makes it visible I can see that the only thing > which is growing in numbers of instances are childConstraintInfo > objects (256 instances per each call). When I hide the panel using > removePopup() - the number of childConstraintInstances does not > decrease. But if I click on the GC button - it goes back to 256 > instances (the panel still exists). > I am 100% sure that none of our custom objects are being created > during the process of popup creation. > > Of course Profiler does not show all the objects (I use standard > filter on flash.*), but if I turn it off, Profiler just dies on the > 4Gb Quad-core machine. And I am unable to turn on the allocation trace > by the same reason - it just kills JVM. > > So we have to find a workaround for this problem. Simplification of > the panel is not an option so far - for complex tasks you have to use > complex tools, hammer & axe are not always the ultimate answer :) > > PS I know that we already reached some limits in our project which are > unknown probably to 99% of developers and most sadly unknown to the > Flex Team. > > Here is the example: > http://bugs.adobe.com/jira/browse/FB-12492 > This popup window has too many mxml entries (due to the states), so it > takes more than 90 seconds to compare it with the previous revision. I > guess that, if people were creating complex interfaces in enterprise > environment, there will be a lot of fuss around this stupid Java > compare tool. > > We also have 4 SWC objects embedded in that panel (graphical > components like buttons or icons developed in Flash). And the Flex > Builder does not like it either - it displays the SWC objects in > Design View only sometimes depending on its own mood and wind > direction. I am unable even to submit a bug on it, as all our small > test cases work fine, the problems appear only in the project environment. > > Cheers, > Dmitri. > > --- In flexcoders@yahoogroups.com<mailto:flexcoders%40yahoogroups.com>, "Ryan > Graham" <Ryan.Graham@> wrote: > > > > > > Hmm... at this point I suppose it depends on what you are doing with > > your popup component. We'd have to see some code to try and pinpoint > > memory-leaks. > > > > > > > > Are you creating a new popup everytime the user requests it, or do you > > create one instance and use the popup manager to show and hide that > > instance, setting data properties accordingly? Are you adding listeners > > to descendant components in the popup? If so are you using weak > > references? For example: > > > > > > > > addEventListener(MouseEvent.CLICK, onClick, false, 0, true); > > > > > > > > and not > > > > > > > > addEventListener(MouseEvent.CLICK, onClick) > > > > > > > > HTH, > > > > Ryan > > > > > > > > From: flexcoders@yahoogroups.com<mailto:flexcoders%40yahoogroups.com> > > [mailto:flexcoders@yahoogroups.com<mailto:flexcoders%40yahoogroups.com>] On > > Behalf Of Dmitri Girski > > Sent: Monday, February 09, 2009 5:37 PM > > To: flexcoders@yahoogroups.com<mailto:flexcoders%40yahoogroups.com> > > Subject: [flexcoders] Re: Memory consumption by the browser > > > > > > > > Hi Ryan, > > > > I tested it on IE7, Chrome, Firefox3, Safari (Mac). They all perform > > in consistent manner, Safari is less greedy, IE & FF are the most > > hungry ones. > > > > But as I said before, I can't see this footprint shrink when I > > minimise browsers. I even set this secret variable for FF - > > trim_on_minimise - no effect. How come? > > > > Cheers, > > Dmitri. > > > > --- In flexcoders@yahoogroups.com<mailto:flexcoders%40yahoogroups.com> > > <mailto:flexcoders%40yahoogroups.com> > > , "Ryan Graham" <Ryan.Graham@> wrote: > > > > > > > > > Is this in all browsers? I know IE 6 (maybe even 7?) has issues with > > > memory management. I consistently see memory leak-like behavior with > > IE > > > 6, for example: > > > > > > > > > > > > Start debug session for basic app - IE reports 60+ MB memory usage > > with > > > a value that steadily climbs as I use the app. > > > > > > Minimize the browser, then bring it back to view - IE memory usage > > drops > > > to ~13 MB. > > > > > > Use the app for a few more minutes, memory footprint grows in size a > > few > > > MB. > > > > > > Minimize, then restore - IE footprint back down to ~13 MB or so. > > > > > > > > > > > > I don't see the same behavior in FF, which is good. The issue might be > > > with memory management within that particular browser, and not the > > Flash > > > Player. > > > > > > > > > > > > HTH, > > > > > > Ryan > > > > > > > > > > > > From: flexcoders@yahoogroups.com<mailto:flexcoders%40yahoogroups.com> > > > <mailto:flexcoders%40yahoogroups.com> > > [mailto:flexcoders@yahoogroups.com<mailto:flexcoders%40yahoogroups.com> > > <mailto:flexcoders%40yahoogroups.com> > > ] On > > > Behalf Of Dmitri Girski > > > Sent: Sunday, February 08, 2009 11:33 PM > > > To: flexcoders@yahoogroups.com<mailto:flexcoders%40yahoogroups.com> > > > <mailto:flexcoders%40yahoogroups.com> > > > Subject: [flexcoders] Memory consumption by the browser > > > > > > > > > > > > Hi everybody, > > > > > > I am trying to solve the problem when my application consumes a lot of > > > memory. > > > When application starts in the browser, Windows/Mac reports the size > > > of browser process around 150Mb. > > > Then, if you start opening/closing the same popup window (which is > > > actually being cached by application), the browser starts to grow. > > > Each appearance of popup consumes around 1.5Mb [sic!] (the consumer is > > > childConstraintInfo, around 256 instances) > > > When you close the pop up, memory sometimes goes down, but never to > > > the original value. Never ever. At least 100-200kb more at a time. > > > > > > OK, when I use a profiler, it shows that application consumes 42Mb. > > > Profiler shows no leaks, even if the number of instances are growing, > > > when I click on GarbageCollector icon it always goes to the original > > > figures, including childConstraintInfo instances. Total Memory stays > > > at 42Mb mark. Good. > > > But the problem is that browser's footprint is still growing. After a > > > couple of hours of work with application it grows to 500-600Mb. > > > > > > In profiler everything looks perfect. In reality - 600Mb app on 1Gb > > > system makes it dead. > > > > > > The question is, what makes browser to be so memory hungry? Why it > > > consumes more & more memory without any attempt to release it? > > > > > > The simplest action with panel instance addPopUp/removePopUp consumes > > > 1.5Mb of memory - I don't believe it. Well, the panels has 3 states > > > and complex layout with ~100 elements. But what makes it so different > > > to the "Hello, World" popup, which seems to be right? > > > > > > Thank you for any thought in advance. > > > > > > Regards, > > > Dmitri. > > > > > > > > > > > > > > > > > > > > > > > > This message is private and confidential. If you have received it in > > error, please notify the sender and remove it from your system. > > > > > > > > > > > > > > > This message is private and confidential. If you have received it in > error, please notify the sender and remove it from your system. > > >
